The data in this Goldman Sachs report is dated (2003) [PDF], but it provides an interesting historical look at what we thought Linux would do back in 2003, and what it actually has done. The good news: the data is off. Goldman Sachs underestimated just how prevalent Linux would become in just four short years. For example, Goldman Sachs' survey revealed that only 39% of enterprise IT respondents were using Linux, but today that number is ~100%. Open source does quite well when the customer gets to vote.
It's going to be a long decade for Microsoft and the UNIX vendors especially because, as PJ at Groklaw notes, the widespread adoption in enterprise IT is only one instance of Linux's growing dominance. Significant, but it's the global tidal wave that should be of more concern to the proprietary vendors. Perhaps this is one reason for Microsoft's patent noise?
Linux-on-Intel appears likely to emerge as the dominant platform in corporate data centers....
In our view, Linux has evolved into an enterprise-class operating system that will have a significant and lasting presence in the IT landscape, and its continued emergence will cause considerable changes in the enterprise IT vendor ecosystem. We believe its strongest effects will be seen in the corporate data center, where we see a shift occurring toward Linux-on-Intel servers away from the current paradigm of proprietary Unix-on-RISC systems. This paradigm shift should have significant implications for the enterprise computing market and for a broad range of vendors in both hardware and software.
Just take a look at that chart above. As the report notes, "IDC forecasts a worldwide server system shipment CAGR of 13.9% through 2006 (from 4.3 million shipments in 2001 to 8.2 million shipments in 2006), but estimates that shipments of servers running Linux will grow at a 33.6% CAGR (from 11.4% of total server shipments in 2001 to 25.2% in 2006), with all other classes of operating systems declining in share." The numbers have been better than projected, though Microsoft continues to grow. Perhaps that is one thing that Goldman Sachs missed: Microsoft is not (yet?) the loser in this competition. UNIX is. Microsoft and the Linux community have yet to truly compete head to head. It's coming, but not yet here.
Interestingly, Goldman Sachs saw Linux's enterprise dominance proceeding at a fairly measured pace. I say "interestingly" because this is precisely what I'm seeing in the open source enterprise applications market. Companies tend to add open source to their existing, proprietary solutions, only displacing (i.e., rip and replace) proprietary applications once the open source application hits a tipping/saturation point.
We believe Linux?s emergence is likely to follow the more measured pace of server hardware upgrade cycles and will not occur within the short timeframe many envision. However, trends such as server consolidation and wider deployment of J2EE-based computing models could significantly accelerate rates of adoption, while slow support from packaged application vendors could be the key dragon the timeframe for Linux's continued emergence.
Why the rapid adoption? Well, despite some opinion to the contrary, it's because Linux is higher-performance, lower-cost, and more secure. This was Goldman Sachs' opinion back in 2003, and customers have been voting for this position with their wallets ever since.
Truth happens, fortunately for open source and unfortunately for the proprietary vendors. Today Linux, tomorrow the world.
Note: I had to update this post when I saw that the report was from 2003. (Missed that on my first read.) As I looked at more current data, however, I found that the post didn't need to be dumped, but rather updated to reflect the fact that Goldman Sachs' suggestion to "fear the penguin" is even more salient (and demonstrated) today than it was back then.